NetXMS Forum

Update successful.
(well, apart from that all links between nodes in manual maps disappeared  :D)

Once more thanks for the help.

Thanks alot, as usual spot on - all problems solved ! (keeping tu actual update until tommorow)

Hello, I have troubles following the wiki page Using APT Repository.

Here are the steps I took so far:
* updated /etc/apt/sources.list file with actual repository address
deb http://packages.netxms.org/debian/ wheezy main

* initialized GnuPG (if I didnt do this - key would not install and apt-get update would throw key errors)
gpg --list-keys
gpg: directory `/root/.gnupg' created
gpg: new configuration file `/root/.gnupg/gpg.conf' created
gpg: WARNING: options in `/root/.gnupg/gpg.conf' are not yet active during this run
gpg: keyring `/root/.gnupg/pubring.gpg' created
gpg: /root/.gnupg/trustdb.gpg: trustdb created
gpg --list-keys - this time no response, suppose that means OK

* download and install key
wget -q -O - http://packages.netxms.org/netxms.gpg | apt-key add -
even tried to just wget http://packages.netxms.org/netxms.gpg and apt-key add netxms.gpg

The problems are these:

1. gpg --list-keys still returns no information, I suppose it should say /root/.gnupg/netxms.gpg ?

2. apt-get update returns less key errors but also less informative...
   Hit http://ftp.cz.debian.org wheezy Release.gpg
   Hit http://ftp.cz.debian.org wheezy-updates Release.gpg
   Hit http://ftp.cz.debian.org wheezy Release
   Hit http://ftp.cz.debian.org wheezy-updates Release
   Hit http://packages.netxms.org wheezy Release.gpg
   Hit http://packages.netxms.org wheezy Release
   Hit http://ftp.cz.debian.org wheezy/main Sources
   Hit http://ftp.cz.debian.org wheezy/main amd64 Packages
   Hit http://security.debian.org wheezy/updates Release.gpg
   Hit http://ftp.cz.debian.org wheezy/main Translation-en
   Hit http://ftp.cz.debian.org wheezy-updates/main Sources
   Hit http://ftp.cz.debian.org wheezy-updates/main amd64 Packages/DiffIndex
   Hit http://ftp.cz.debian.org wheezy-updates/main Translation-en/DiffIndex
   Hit http://security.debian.org wheezy/updates Release
   Hit http://packages.netxms.org wheezy/main amd64 Packages
   Hit http://security.debian.org wheezy/updates/main Sources
   Hit http://security.debian.org wheezy/updates/main amd64 Packages
   Hit http://security.debian.org wheezy/updates/main Translation-en
   Ign http://packages.netxms.org wheezy/main Translation-en_US
   Ign http://packages.netxms.org wheezy/main Translation-en
   Reading package lists... Done
   W: There is no public key available for the following key IDs:
   7638D0442B90D010
   W: There is no public key available for the following key IDs:
   7638D0442B90D010
   W: There is no public key available for the following key IDs:
   9D6D8F6BC857C906

Not really sure if this is connected with netxms.gpg key at all.

3. how do I check if the key is installed correctly ?

4. is apt-get install netxms=1.2.17  the  correct command to update to 1.2.17 version or should I instruct to update all the packages specifically (netxms-agent netxms-server netxms-server-mysql netxms-base) ?
   Not sure if there are versions for all packages or not. Our current version is 1.2.16.

With regards

Milan Sperl

Thanks, one last question (I know it is not about Debian anymore but I do not want to create unnecessary topics, or should I?).

Is it possible to create standalone page showing one dashboard ?

• either by autologin pointing directly to the needed dashboard
• or by configuring webpage itself to display dashboard using present Tomcat
  (though I am not sure if I can do such a thing and definitely need some pointers...)

I was kind of hoping something like this was already discussed but did not find any clue on the forum.

With regards

Milan Sperl

Hello, just to close this topic:

tried several scenarios with importing fewer tables, or without the xdata_

• ended up with various levels of "desired result", all lacking (mainly missing functional maps)

about the "tricky part"... was not worth the hassle 

• simply recreated everything from scratch

Now we have nice and tidy fresh setup (hope it stays that way).

Just a question regarding system requirements - previous installation had 2G RAM, new one has only 1G. Would you suggest adding more ?
Considering on both setups we had (have) "problems" with Tomcat webservice hanging now and then (when it literally inflates ram usage (and eventually hits swap file) for no reason).

Right now NetXMS stats with final number of objects:
Total number of objects:    23413
Number of monitored nodes:  918
Number of collectable DCIs:  928

CPU E31220 @ 3.10GHz

Memory at the moment:
MemTotal:        1027020 kB
MemFree:          69964 kB
Buffers:          13112 kB
Cached:          181620 kB
SwapCached:        11516 kB

Using virtual server, right now I'm considering adding up to more than 2GB RAM, not sure if it helps or not though.

With regards,

Milan Sperl

Neat! Thanks for pointing that out for me.
This is perfect solution.

Have to apologize for not using the search button (really, when one does not use search it turns out to be really easy problem which has already been solved and well documented  :o).

Hello,

  I have Infrastructure Services structure built up (for dependency script, info on hierarchy in messages and to help build up network maps). And I was wondering if there is anything I can do to check if there are new nodes that need to be "bound" to the structure.

I was thinking of building a script that gets node on a discovery poll and compares if it is somewhere in Infrastructure Services. If not, add it to some temporary folder in the infrastructure (so it is "found" on the next poll, and highlighted as "new").
Problems I can see are load it could make on every poll, but more importantly - I don't have idea how to check the infrastructure to compare it. The node can be anywhere deep in the infrastructure, how would you solve such check ?

Script that does it once per request might be even better and this has led me to another question - is there some kind of parameter in database that might say there is a link and the node has been "bound" somewhere ?
For start it would be sufficient to run a query against the DB to check if nodes are "not bound anywhere".
Have not found suitable parameter though.

With regards,

M.Sperl

Not sure about the "still scanning outside" part. It makes sense though - active discovery scans the ranges given in Active Discovery Targets, passive discovery adds some more, then Filter kicks in and has a final say as to what node will be added or not.

With this setting it seems to behave a little better - no unwanted nodes appear in the database.

We are still having Entire Network show much more subnets just because some devices are "somehow" visible from the unwanted subnets.
It would be really great if subnets in the Entire Network sections followed the same rules.

Thanks for the clarification.

So the only option for having "tight" subnets is set active-passive discovery and set what subnets should be scanned. And to prevent the "passive" part going outside the active boundaries the filter needs to be specified with the same set as the active list of subnets.

Hello, once again I have been reviewing Network Discovery Configuration and found out that it is not working as I need it to.

What I want to achieve:

• scan only subnets that are precisely specified

At first I set:

• General - network discovery setting - Active and passive
• Default SNMP community string - filled (our own string)
• Active Discovery Targets - subnets specified
• Filter - No filtering

But this scanned literally everything.

I am not sure if setting General to Disabled will help (according to wiki the list in Active Discovery Targets has no effect if active discovery is off).
Seeing as General has three options:

• Disabled
• Passive
• Active and passive
  I am not really sure if "Disabled" means "Active only" or "Nothing at all".

Or should I check Active and passive and add the same list of desired subnets to the Address Filters ?

Clarification would be really helpful.

With regards,

Milan Sperl

Hello, thanks for the clarifications.

As for

1. I used "apt-get install netxms-agent netxms-server netxms-server-mysql netxms-base", and it installed nicely with agent and mysql backend so my question was if the options are needed at all.
   SOLVED

2. yes, that was the question - deb packages were missing sample config files so I copied the ones I had from previous instalation and the question was if the config files are persistent so the copy is safe way of getting it right. It works so I suppose it is safe
   SOLVED

3. yes, we have full backup of old DB with MySQL Admin which has very simple output which allows us to select parts to restore, there should not be problem with selecting and restoring parts of the DB

• the question is if we can get to migrate "just" hierarchy and network maps
• as these are dependent on IDs of discovered nodes this all boils down to - how to migrate discovered nodes
• everything else is nicely visible in the DB and can be "copy-pasted" many ways, there is no problem migrating scripts, events, rules, filters, templates...
• I would like to avoid migrating whole database for the purpose of "fresh" start as I think the old one might have too many unnecessary parts which caused troubles in the past

That should get us started on new machine all over again.

Thanks for your support, very appreciated.
With regards,

Milan Sperl

Hello, after some reconsiderations we decided to try installing NetXMS on Debian from packages.

• made fresh Debian installation
• updated apt, dist-upbrade, configured ssh, iptables
• installed mysql and set up DB for NetXMS
• added "deb https://www.netxms.org/apt wheezy main" to sources.list
• tried to install NetXMS (apt-get install netxms-agent netxms-server netxms-server-mysql netxms-base)

And this is where my problems start, I am not fully used to debian and packages installation so this is where the problem might lie but I have got questions that need to be answered:

1. how can I add options to apt-get install command to specify the configuration ? (is it needed ?)
   --with-server --with-mysql (--with-agent) --prefix=/opt/netxms
   (never added the --with-agent option, is it really needed ? always had the agent anyway)

2. where can I get sample config files ? (I suppose these should be present in contrib directory, but there is no such)

• is copy from old installation safe option ? (without fear something was changed in newer versions)

3. is it safe to migrate only some DBs from old installation if we want to keep only list of found IPs with old IDs so that we can later import old network maps and hierarchy structure for example ?

Can you please help me by answering these questions?
(the third one is not mandatory, will experiment on it as soon as I get the server running)

With regards,

Milan Sperl

edit: the server is up and running, questions still stand (are there options available when installing via packages?, where to get up-to-date config files and how to migrate only infrastructure + network maps)

p.s. for testing purposes I need block all netxms comunication going outside, is it sufficient to block only outgoing ports 161 and 162 ?

When you remove those DCIs with zero values manually, they pop back up ?
If not, it must have been something while creating the DCI (mind that whenever you open or modify the DCI it reapplies).
If these return after modyfying/opening the root DCI, there is something in it.

Hello, sorry for the delay.

Its my first time debugging, used http://www.ibm.com/developerworks/linux/library/l-gdb/ as reference.
Had to enter "file /opt/netxms/bin/netxmsd" into gdb before "run -D5".

Debug finished with

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffff0132700 (LWP 5879)]
AlarmManager::watchdogThread (this=0x7ffff7dd7e80) at alarm.cpp:1105
1105                            if ((m_pAlarmList[i].ackTimeout != 0) &&

bt output

#0  AlarmManager::watchdogThread (this=0x7ffff7dd7e80) at alarm.cpp:1105
#1  0x00007ffff7aea849 in WatchdogThreadStarter (pArg=<value optimized out>) at alarm.cpp:204
#2  0x00007ffff6137851 in start_thread (arg=0x7ffff0132700) at pthread_create.c:301
#3  0x00007ffff534b90d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115

With regards,
Milan Sperl

Yes,

  installed missing flex and bison - non development versions (how come we did not need these before?).
Downloaded netxms-1.2.14.tar.gz, extracted, swapped parser.y, configured install, make, make install, started agent, server... same results - huge amount of warnings because of filters, tried to switch them off but server fails again

Hello, we are running CentOS release 6.4 (final) version: 2.6.32-358.14.1.el6.x86_64.

It is 64 bit server.

Hello, what was the solution?

I have got problems with upgrade from 1.2.12 to 1.2.14 - server shuts down after a while.
It all happens while trying to manage things - 2 times trying upload image, once opening scripts, once opening network map. These actions actually process but server shuts down after few seconds.

The only reasonable log from debug is

[27-May-2014 16:48:31.062] [WARN ] Failed to execute filter script for template object 26027 "H3C (-25506-,sys65) - filtr ok - bez DCI": Error 11 in line 13: Function not found
[27-May-2014 16:48:31.062] [WARN ] Failed to execute filter script for template object 26029 "H3C (-25506-,sys47) - filtr ok - bez DCI": Error 11 in line 4: Function not found
[27-May-2014 16:48:34.473] [WARN ] Failed to execute filter script for template object 26027 "H3C (-25506-,sys65) - filtr ok - bez DCI": Error 11 in line 13: Function not found
[27-May-2014 16:48:34.474] [WARN ] Failed to execute filter script for template object 26029 "H3C (-25506-,sys47) - filtr ok - bez DCI": Error 11 in line 4: Function not found
[27-May-2014 16:48:50.606] [WARN ] Failed to execute filter script for template object 26027 "H3C (-25506-,sys65) - filtr ok - bez DCI": Error 11 in line 13: Function not found
[27-May-2014 16:48:50.607] [WARN ] Failed to execute filter script for template object 26029 "H3C (-25506-,sys47) - filtr ok - bez DCI": Error 11 in line 4: Function not found
Segmentation fault

There seems to be some problems with scripts and filters (which were ok in 1.2.12).

Tried to search for the segmentation fault problem on forum but found nothing definite...

Attaching whole log.

---

netxms-debug.txt

The filter discussed above is for the instanced DCIs, right ?
At the time of applying of the DCI it checks if there are values (non-zero and non-null) and creates real DCI for that one OID.

Where exactly are you getting those 0 values ?

If there are semi-stable values in the OIDs and DCIs are created at that time, these will exist even if OID return 0 afterwards... Check history of the DCIs which are showing 0 now.

You can also check snmpwalk if the 0 value is correct or not.

Hello, just a question...

IF value is 0, doesnt it automatically mean that return value is equivalent to return false ?
(thus not passing the filter script)

MS

Hi, maybe you could solve it by filters in event processing policy ?

Something along these lines:
IF

• SYS_IF_DOWN
• and:
  parentNode = GetObjectParents($if); // not sure if "$if" (Get ObjectParents reference)
  if ((parentNode->status == 4) return 0;

For more details see my solution of dependency based on infrastructure, if it helps... here

With regards,

MS